priority in origin, takes the ascendancy and appropriates the protoplasm for the development of the sporiferous head.

That the umbilical stalk is formed directly from the stem arising from an anastomosis and is at least the chief channel through which the fructification receives its supply of nourishment is indicated by a number of facts. Its base is at first in connection with the fine anastomosing hyphae, but as the plant reaches its full size it shrivels and usually leaves at maturity little more of a remnant than shown in fig. 9. In cases in which its base remains, it may be seen to be somewhat raised above the attachment of the two pairs of legs, as if it had been torn out of the substratum by the growth of these latter. The legs are obviously a later production, as is seen by fig. 8, where the anchoring rhizoids are just beginning to be given off toward their bases. The legs, as may be seen in the left hind leg of the specimen just mentioned, are supplied with rhizoids which at times certainly are comparatively short and end blindly; although in other cases a few of the rhizoidal branches can be followed for a certain distance into delicate filaments characteristic of the mycelium. This latter condition, however, is not so typical as for the umbilical stalk.

Neither the vegetative mycelium nor the hyphae of the fructification are septate during growth. Septation regularly takes place in the neck, legs, and branches of the head when the protoplasm which they contain becomes used up in spore-formation, and mature spores showing the characteristic echinulations are to be found on fructifications in which no septa have as yet been found. The cross walls are thin in comparison with the thick lateral walls and quite irregularly disposed. Often one connects the lateral wall with an adjacent septum or extends as a shelf but part way across the hypha. The branches of the crown easily separate at their septa, and by judicious tapping with a needle the crown may be denuded of nearly all its sporiferous heads and branches.

The only known form to which Thamnocephalis shows any close relationship is Sigmoidiomyces dispiroides Thaxter, and the two genera evidently form a group by themselves. The method of spore

4 THAXTER, R., North American Hyphomycetes. BOT. GAZETTE 16:22. pl. 4. figs. 15-18. 1891. Figures reproduced in ENGLER and PRANTL'S Pflanzenfamilien 11:427. figs. 220 G-H.

formation on paired heads which are located at points of branching of the fertile hyphae is essentially the same for both species. The outgrowths on the convexities of the sterile branches in Sigmoidiomyces are represented in Thamnocephalis by mere protuberances; the spores and their echinulations are much reduced in the latter; and the branches of the fertile hyphae are not markedly curved in the mature condition. Sigmoideomyces fails, moreover, to show that striking differentiation into fertile crown, main stalk, and rhizoidal props, which is a unique feature of Thamnocephalis. Sigmoideomyces was described from mature material, and it was not possible to determine the nature of the mycelium nor the time of septation of the fertile hyphae, but the similarity in the two forms would suggest that much the same condition exists as in Thamnocephalis.

MATRUCHOT has emphasized the systematic importance of nonseptate hyphae, and the condition in Cunninghamella shows that the presence of sporangia is not a sine qua non for admission to the company of the mucors. The absence of septa in young growing hyphae is a mucor character, but hardly less characteristic is their presence in varying abundance in older hyphae from which the protoplasm has been withdrawn. Cunninghamella is no exception to this statement (fig. 3), but perhaps Piptocephalis and Spinellus macrocarpus furnish as striking examples of septation in fertile hyphae as the group affords. The delicate anastomosing mycelial hyphae further remind one of the similar condition in Syncephalis. THAXTER (l. c.), MATRUCHOT (1. c.), and others have recognized the possibility of Rhopalomyces belonging to the Mucorineae. Without evidence from cultures. the relationship of Sigmoideomyces and Thamnocephalis with the mucors must be left as only a suggestion.

Microscopic preparations and herbarium material of Thamnocephalis quadrupedata have been deposited in the Cryptogamic Herbarium of Harvard University.

NAPLES BIOLOGICAL STATION.

EXPLANATION OF PLATE VI.

The drawings were outlined with the aid of a camera lucida with the combination of Leitz and Bausch & Lomb lenses noted and have been reduced about one-quarter in reproduction.

Cunninghamella echinulata Thaxter.

FIG. 1. Development of progametes; obj. 7. oc. 3.

FIG. 2. Abstriction of gametes; obj. 7. oc. 3.

FIG. 3. Young zygote in connection with short conidiophore through hypha which has lost its protoplasm and become septate; obj. 7. oc. 3.

FIG. 4. Mature zygospore; obj. 7. oc. 3.

FIG. 5. "Twinned" zygotes; obj. 7. oc. 3.

Thamnocephalis quadrupedata, n. sp.

FIG. 6. Hyphae of mycelium showing anastomoses; obj. 7. oc. 3.

FIG. 7. Early stages of development of fructification; obj. 7. oc. 1.

FIG. 8. Young fructification showing basal portion and formation of sporiferous heads at first node of the crown; obj. 7. OC. I.

FIG. 9. Mature fructification from which many of the fertile branches and spores have been shed; obj. D. oc. 2.

FIG. 10. Last fertile node of crown showing sterile terminal branches and young sporiferous heads from which the spores have not yet developed; obj. .

OC. 3.

FIG. 11. Fifth fertile node with sporiferous heads and spores; obj. 2. oc. 3.